makedensitybasedclusterer.java

数据挖掘中聚类的算法

    return n;
  }

  /**
   * Computes the log of the conditional density (per cluster) for a given instance.
   * 
   * @param inst the instance to compute the density for
   * @return an array containing the estimated densities
   * @throws Exception if the density could not be computed
   * successfully
   */
  public double[] logDensityPerClusterForInstance(Instance inst) throws Exception {

    int i, j;
    double logprob;
    double[] wghts = new double[m_wrappedClusterer.numberOfClusters()];
    
    m_replaceMissing.input(inst);
    inst = m_replaceMissing.output();

    for (i = 0; i < m_wrappedClusterer.numberOfClusters(); i++) {
      logprob = 0;
      for (j = 0; j < inst.numAttributes(); j++) {
	if (!inst.isMissing(j)) {
	  if (inst.attribute(j).isNominal()) {
	    logprob += Math.log(m_model[i][j].getProbability(inst.value(j)));
	  } else { // numeric attribute
	    logprob += logNormalDens(inst.value(j), 
				     m_modelNormal[i][j][0], 
				     m_modelNormal[i][j][1]);
	  }
	}
      }
      wghts[i] = logprob;
    }
    return  wghts;
  }

  /** Constant for normal distribution. */
  private static double m_normConst = 0.5 * Math.log(2 * Math.PI);

  /**
   * Density function of normal distribution.
   * @param x input value
   * @param mean mean of distribution
   * @param stdDev standard deviation of distribution
   * @return the density
   */
  private double logNormalDens (double x, double mean, double stdDev) {

    double diff = x - mean;
    
    return - (diff * diff / (2 * stdDev * stdDev))  - m_normConst - Math.log(stdDev);
  }
  
  /**
   * Returns the number of clusters.
   *
   * @return the number of clusters generated for a training dataset.
   * @throws Exception if number of clusters could not be returned successfully
   */
  public int numberOfClusters() throws Exception {

    return m_wrappedClusterer.numberOfClusters();
  }

  /**
   * Returns a description of the clusterer.
   *
   * @return a string containing a description of the clusterer
   */
  public String toString() {
    StringBuffer text = new StringBuffer();
    text.append("MakeDensityBasedClusterer: \n\nWrapped clusterer: " 
		+ m_wrappedClusterer.toString());

    text.append("\nFitted estimators (with ML estimates of variance):\n");
    
    for (int j = 0; j < m_priors.length; j++) {
      text.append("\nCluster: " + j + " Prior probability: " 
		  + Utils.doubleToString(m_priors[j], 4) + "\n\n");
      
      for (int i = 0; i < m_model[0].length; i++) {
        text.append("Attribute: " + m_theInstances.attribute(i).name() + "\n");
	
        if (m_theInstances.attribute(i).isNominal()) {
          if (m_model[j][i] != null) {
            text.append(m_model[j][i].toString());
          }
        }
        else {
          text.append("Normal Distribution. Mean = " 
		      + Utils.doubleToString(m_modelNormal[j][i][0], 4) 
		      + " StdDev = " 
		      + Utils.doubleToString(m_modelNormal[j][i][1], 4) 
		      + "\n");
        }
      }
    }

    return  text.toString();
  }
  
  /**
   * Returns the tip text for this property
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String clustererTipText() {
    return "the clusterer to wrap";
  }

  /**
   * Sets the clusterer to wrap.
   *
   * @param toWrap the clusterer
   */
  public void setClusterer(Clusterer toWrap) {

    m_wrappedClusterer = toWrap;
  }

  /**
   * Gets the clusterer being wrapped.
   *
   * @return the clusterer
   */
  public Clusterer getClusterer() {

    return m_wrappedClusterer;
  }
  
  /**
   * Returns the tip text for this property
   * @return tip text for this property suitable for
   * displaying in the explorer/experimenter gui
   */
  public String minStdDevTipText() {
    return "set minimum allowable standard deviation";
  }

  /**
   * Set the minimum value for standard deviation when calculating
   * normal density. Reducing this value can help prevent arithmetic
   * overflow resulting from multiplying large densities (arising from small
   * standard deviations) when there are many singleton or near singleton
   * values.
   * @param m minimum value for standard deviation
   */
  public void setMinStdDev(double m) {
    m_minStdDev = m;
  }

  /**
   * Get the minimum allowable standard deviation.
   * @return the minumum allowable standard deviation
   */
  public double getMinStdDev() {
    return m_minStdDev;
  }

  /**
   * Returns an enumeration describing the available options..
   *
   * @return an enumeration of all the available options.
   */
  public Enumeration listOptions() {
    Vector result = new Vector();

    result.addElement(new Option(
	"\tminimum allowable standard deviation for normal density computation "
	+"\n\t(default 1e-6)"
	,"M",1,"-M <num>"));
	
    result.addElement(new Option(
	"\tClusterer to wrap.\n"
	+ "\t(default " + defaultClustererString() + ")",
	"W", 1,"-W <clusterer name>"));

    if ((m_wrappedClusterer != null) &&
	(m_wrappedClusterer instanceof OptionHandler)) {
      result.addElement(new Option(
	  "",
	  "", 0, "\nOptions specific to clusterer "
	  + m_wrappedClusterer.getClass().getName() + ":"));
      Enumeration enu = ((OptionHandler)m_wrappedClusterer).listOptions();
      while (enu.hasMoreElements()) {
	result.addElement(enu.nextElement());
      }
    }
    
    return result.elements();
  }

  /**
   * Parses a given list of options. <p/>
   *
   <!-- options-start -->
   * Valid options are: <p/>
   * 
   * <pre> -M &lt;num&gt;
   *  minimum allowable standard deviation for normal density computation 
   *  (default 1e-6)</pre>
   * 
   * <pre> -W &lt;clusterer name&gt;
   *  Clusterer to wrap.
   *  (default weka.clusterers.SimpleKMeans)</pre>
   * 
   * <pre> 
   * Options specific to clusterer weka.clusterers.SimpleKMeans:
   * </pre>
   * 
   * <pre> -N &lt;num&gt;
   *  number of clusters. (default = 2).</pre>
   * 
   * <pre> -S &lt;num&gt;
   *  random number seed.
   *  (default 10)</pre>
   * 
   <!-- options-end -->
   *
   * @param options the list of options as an array of strings
   * @throws Exception if an option is not supported
   */
  public void setOptions(String[] options) throws Exception {

    String optionString = Utils.getOption('M', options);
    if (optionString.length() != 0)
      setMinStdDev((new Double(optionString)).doubleValue());
    else
      setMinStdDev(1e-6);
     
    String wString = Utils.getOption('W', options);
    if (wString.length() == 0)
      wString = defaultClustererString();
    setClusterer(Clusterer.forName(wString, Utils.partitionOptions(options)));
  }

  /**
   * Gets the current settings of the clusterer.
   *
   * @return an array of strings suitable for passing to setOptions()
   */
  public String[] getOptions() {

    String [] clustererOptions = new String [0];
    if ((m_wrappedClusterer != null) &&
	(m_wrappedClusterer instanceof OptionHandler)) {
      clustererOptions = ((OptionHandler)m_wrappedClusterer).getOptions();
    }
    String [] options = new String [clustererOptions.length + 5];
    int current = 0;

    options[current++] = "-M";
    options[current++] = ""+getMinStdDev();

    if (getClusterer() != null) {
      options[current++] = "-W";
      options[current++] = getClusterer().getClass().getName();
    }
    options[current++] = "--";

    System.arraycopy(clustererOptions, 0, options, current, 
		     clustererOptions.length);
    current += clustererOptions.length;
    while (current < options.length) {
      options[current++] = "";
    }
    return options;
  }

  /**
   * Main method for testing this class.
   *
   * @param argv the options
   */
  public static void main(String [] argv) {
    runClusterer(new MakeDensityBasedClusterer(), argv);
  }
}